基于大数据分析的犯罪模式挖掘与犯罪预测研究

项目名称： 基于大数据分析的犯罪模式挖掘与犯罪预测研究

项目编号： No.91546103

项目类型： 重大研究计划

立项/批准年度： 2016

项目学科： 宏观管理与政策

项目作者： 潘李伟

作者单位： 中国电子科技集团公司第三十八研究所

项目金额： 38万元

中文摘要： 本课题以新疆暴力恐怖犯罪为背景研究犯罪分析问题。以事后分析为主的传统警务使用有限的犯罪数据、破案率低，分析方法缺乏准确性和时效性，尤其不能满足实时性要求较高、需要提前预警的暴力恐怖型犯罪。为此，本课题提出基于大数据的犯罪模式挖掘与犯罪预测，通过情报收集、数据融合、数据挖掘及社会网络分析等技术，融合异构数据源修正、扩展和完善犯罪数据，以客观全面地识别犯罪模式、预测未来犯罪。本课题：(1)提出基于大数据的犯罪模式挖掘与犯罪预测框架，研究数据融合方法；(2)研究犯罪时空模式、个体犯罪模式和群体犯罪模式的挖掘方法，以及犯罪预测等重要内容；(3)实现基于反恐怖情报信息的大数据犯罪分析平台，完善新疆地区的反恐维稳工作。本课题通过大数据犯罪分析，实现数据驱动的预测性警务以有效协同和补充传统警务；通过新疆的试点部署，降低新疆乃至全国的犯罪率，维护公共安全和社会安定，因此具有重要的理论意义和应用价值。

中文关键词： 暴恐犯罪；犯罪分析；预测性警务；数据融合；数据挖掘

英文摘要： This project aims to research on crime data analysis with violent terrorist crime in Xinjiang province as the background. Traditional postmortem analysis led policing typically employs limited crime data with low rate of solved criminal cases. Besides, the accuracy and time-efficiency of traditional crime analysis methods are not satisfied enough. Especially, traditional policing fails for violent terrorist crime, which has high real-time demand and requires warning and precaution mechanism. To this end, this project proposes crime pattern mining and crime forecasting using big data analytics. Leveraging techniques in intelligence collection, data fusion, data mining and social network analysis, multiple heterogeneous data sources can be fused and integrated to amend, expand and improve basic crime data, so that crime patterns can be fully identified, and further crime can be predicted in the right perspective. Specifically, we: (1) propose the framework of big data analytics oriented for crime pattern mining and crime forecasting, and research on data fusion methods; (2) research on crime spatial and temporal pattern, individual crime pattern and group crime pattern, as well as crime prediction modeling; (3) implement the big data crime analytics platform based on counter-terrorism intelligence information in Xinjiang, to replenish the efforts in counter-terrorism and safeguard stability. Through the big data crime analysis research, this project aims to achieve data-driven predictive policing to collaborate and complement traditional policing. Also, through the experimental deployment in Xinjiang, this project would reduce the crime rate in Xinjiang and even the national wide. Therefore, this project has profound influences in public safety and social stability, and thus has important theoretical significance and application value.

英文关键词： Violent Terrorist Crime;Crime Analysis;Predictive Policing;Data Fusion;Data Mining